1. Field of the Invention
This invention relates generally to time-resolved recording of three or more optical parameters simultaneously; and more particularly to novel methods and apparatus for making such measurements on an extremely short time scale, using lidar or a streak tube—or related technologies such as lenslet arrays—or combinations of these. Certain forms of the invention enable provision of a compact single-laser-pulse scannerless multidimensional imaging system using plural-slit streak-tube imaging lidar or “PSSTIL”. The system is also capable of making plural-wavelength-band spectrally discriminating recordings of objects or phenomena, as well as plural-polarization-state recordings, and also combinations of such novel measurements.
2. Related Art
(a) Conventional streak lidar—The term “lidar” (in English pronounced “LIE-dahr”), by analogy to “radar”, means “light detection and ranging”. The use of lidar is greatly enhanced by incorporating a streak tube—an electrooptical system for time resolving lidar returns to a remarkably fine degree.
Several advanced forms of the streak-tube imaging lidar or “STIL” technology are presented in other patent documents. See, for example, U.S. Pat. No. 5,467,122 and PCT publication WO 98/10372.
BENEFITS: Many strengths of a conventional STIL system appear in Table 1, and most of these are discussed in more detail in this section. This technology has demonstrated the capability to collect range (or other time-related) information with dynamic range and bandwidth that cannot be achieved using earlier conventional signal-digitization electronics.
TABLE 1Benefits of the conventional streak-tubeimaging lidar (STIL) approach.FeatureBenefit12-bitAllows robust operations in high-contrast scenes.linear dy-User need not spend time keeping system “centered”namic rangein the dynamic range.No electronic digitization can achieve this withoutsignificant compression (use of a logarithmic am-plifier) that introduces artifacts in data.ControllableCan change digitization rate “on the fly”, whichrange-allows the operator to start out with coarse rangeoperationresolution and “zoom” in on areas of interest.from1 cm range resolution has been demonstrated atd. c. toshort ranges; 15 cm, from an aircraft.multiGHzConventional high-speed digitization electronicsare designed for only one speed.Fast dataSTIL operates on a single short-pulse (<10 nsec)collectiontime-of-flight measurement; thus no long integra-tion or multiple pulsing is necessary.No target distortion/blur from moving source ortarget.CompactThe volume of commercial streak-tube electronicsruggedizedpackaging has been reduced by a factor of twenty.packageRuggedized hardware for helicopter environment hasbeen fabricated.System can be placed on a variety of platforms.HighThe streak tube has a noiseless gain (noise factorinternalis ~1) due to each accelerated photoelectron gener-gainating approximately three hundred photons on thephosphor screen.Higher gain (>104) available using a microchannelplate (MCP) if necessary.Raises small signals above the amplifier noise.Simulta-No need for multiple sensors.neousAllows significant improvements to ATR algorithmsrange andfor shape matching and for clutter reduction.contrastNo registration/scale issues between range and con-imagestrast data.High frameTransmitters and receivers have been demonstratedratesat 400 Hz frame rates-ideal for large-areashownsearching.Ideal for very fast moving targets or sensors.RapidAll processing within a single image frame is con-processingventionally performed with multiple DSP computers.Allows rapid real-time display for an operator.Reduces volume and electrical power requirementsfor computer.
For example, linear twelve-bit dynamic range has been shown at controllable bandwidths up to and beyond 3 GHz. A fundamental advantage of a streak-tube-based lidar system is that it can provide hundreds or even thousands of channels of sampling at more than 3 GHz, with true twelve-bit dynamic range.
Such a receiver system provides range sample “bins” (i.e. discrete range-sampling intervals) that can be as small as five centimeters (two inches) long, and provides 4096 levels of grayscale imagery, both of which are important for robust operations. The small range bins provide optimal ranging capability, and the large dynamic range reduces the effort of trying to keep the scene illumination in the middle of the response curve of a limited-dynamic-range receiver. Such performance has been demonstrated in a laser radar configuration (J. McLean, “High Resolution 3-D Underwater Imaging”, Proc. SPIE 3761, 1999).
BASIC TUBE ARCHITECTURE AND OPERATION: A streak tube (FIG. 1) as conventionally built nowadays is very similar to a standard image-intensifier tube, in that it is an evacuated tube with a photocathode, producing electrons which are accelerated by very high voltages to a phosphor screen. In operation of a typical system, each such electron ejects roughly three hundred photons from the phosphor, which is then collected by an image-recording device such as a CCD.
A major difference is that a streak tube has an extra pair of plates that deflect an electron beam, somewhat as do the deflection plates in an ordinary cathode ray tube (CRT) tube used in most oscilloscopes, television sets and computer monitors. In conventional STIL operation, input photons are limited to a single slit-formed image—causing the electron beam within the tube to be slit-shaped.
A fast ramp voltage is applied to the deflection plates, very rapidly and continuously displacing or “streaking” the slit-shaped electron beam, parallel to its narrow dimension, from the top (as oriented in FIG. 1) of the phosphor screen to the bottom—effectively creating a series of line images formed at different times during the sweep. Thereby time information is impressed upon the screen image in the streak direction (here vertical), while spatial information is arrayed along the slit length.
The array of internal electronic line images in turn constitutes a latent areal image—which can be picked up (“developed”) by phosphor on the screen. Most typically a charge-coupled-device (CCD) camera is attached to the streak tube to collect the image from the phosphor screen.
In this way the image is reconverted to an external electronic image by a CCD. The CCD output is digitized, interpreted, and if desired saved or displayed by receiving electronics.
One of the two dimensions of each two-dimensional image acquired in this way is azimuth (taking the dimension parallel to the long dimension of the slit as extending left and right)—just as with a common photographic or video camera. The other of the two dimensions, however, is unlike what an ordinary camera captures.
More specifically, the STIL images represent azimuth vs. range from the apparatus—not vs. the commonplace orthogonally visible dimension as with a common camera. Thus for example if a two-dimensional image of an ocean volume is acquired by an instrument pointed vertically downward into the sea, the two dimensions are azimuth and ocean depth.
The operation described here should not be confused with that of a so-called “framing camera”, whose tube internal geometry is commonly identical but which usually lacks an optical input slit, and whose deflection system is differently energized—so that more ordinary two-dimensional images of a scene are formed at the phosphor screen. Often such images are sized to fit on just a fraction of the screen area, and the deflection plates quickly step the two-dimensional-image position (in some instruments during a blanking interval), rather than displacing it continuously as in streaking.
BASIC SYSTEM OPERATION: In a typical conventional streak-tube lidar configuration, a short-duration high-energy laser pulse is emitted. The emitted beam is spread out into a single thin, fan-shaped beam or line, which is directed toward a landscape, ocean volume, or other region of interest—and the receiver optics image the line back onto the slit input to the streak tube. (A later portion of this document discusses the phrase “thin, fan-shaped beam” in further detail.)
In such a standard STIL system, coverage of the region of interest in the dimension perpendicular to the line illumination (FIG. 2) is generally accomplished through motion of a vehicle carrying the emitter and sensor successive pulses of the beam. Formation of a complete volumetric image therefore requires a series of pulses, each yielding a respective individual range-vs.-azimuth image.
Taking the laser projection direction as horizontal in FIG. 2(a), the vehicle direction should be vertical—as for instance in a vertically moving helicopter. In this case, each areal screen image represents a horizontal map, at a respective altitude, with the measuring instrument located above the top edge of the map and the remote horizon along the bottom edge.
Alternatively, reverting to the earlier example of a downward-looking instrument over the ocean, vehicle motion should be horizontal. In this case each areal screen image represents a vertical slice of the ocean below the vehicle, at a respective position along the vehicle's horizontal path.
This is sometimes familiarly called a “pushbroom” system. A demonstrated alternative to vehicle-based data acquisition is a one-dimensional scanner system used from a fixed platform.
The deflection system of the streak tube is set to streak the electron beam completely down the phosphor screen in some specific time, called the “sweep time” of the tube. This also corresponds to the total range gate time (i.e., the total amount of time during which the system digitizes range data).
Ordinarily the sweep time is adjusted to fully display some interval of interest for exploring a particular region, as for instance some specific ocean depth from which useful beam return can be obtained—taking into account turbidity of the water. The starting point of the range gate is controlled by the trigger signal used to begin the sweep.
Computer control of both the sweep time and the sweep-start trigger provides the operator a flexible lidar system that can very rapidly change its range-gate size, its range-digitization starting point, and also its range-sampling resolution. This enables the system to search large areas of range with coarse range resolution, and then “zoom in” to obtain a high-resolution image around a discovered region of particular interest. For example, in one pulse the system could capture a range from 5 km to 7 km at low resolution, and then on the next laser pulse zoom in to 6 km±50 m and thereby image an object of prospective interest at the highest resolution.
Each column of CCD pixels corresponds to one channel of digitized range data, such as would be collected from a single time-resolved detector—for instance a photomultiplier tube (PMT) or an avalanche photodiode (APD). Each row is the slit image at a different time.
The size (in units of time) of the previously mentioned range bins is simply the sweep time divided by the number of pixels in the CCD columns. Such values are readily converted into distance units through multiplication by the speed of light in the relevant medium or media.
MODERN-DAY ENHANCEMENTS: Considerable practical advancement is now available in state-of-the-art streak-tube technology. Such advances include a compact ruggedized package suitable for helicopter environment.
Such a unit (FIG. 3) is only about 15 cm (6 inches) wide, 47 cm (19 inches) long, and 37 cm (15 inches) in diameter. This kind of device has complete computer control of all streak-tube parameters, including high-voltage supplies.
Available as well are continuously variable linear sweep speeds from 50 nsec to 2 μsec. High-speed tube gating without a microchannel plate (MCP) is also offered, for enhanced signal-to-noise ratio.
ADVANCED COMMERCIAL FORM: The assignee of this patent document, Aret é Associates (of Sherman Oaks, Calif., and Tucson, Ariz.), has developed an airborne STIL system for bathymetry and terrestrial mapping. This device contains a diode-pumped solid-state Nd:YAG laser that is frequency doubled to 532 nm. This wavelength was chosen for maximum water penetration for the bathymetry task and for proximity to the peak of the streak-tube photocathode responsivity curve.
A raw image frame taken by STIL during airborne terrestrial mapping data collection (FIG. 4[b]) and a volume reconstruction from numerous such frames (FIG. 4[c]) compare interestingly with a conventional photograph (FIG. 4[a]). A like comparison is also shown (FIG. 5) for an object roughly 1 m (39 inches) in diameter and imaged through 6 m (20 feet) of seawater.
In these views, naturally the conventional photo gives a clearer and sharper image. One goal of the STIL imaging, however, is to obtain images and reconstructions under circumstances that preclude effective use of ordinary photos.
Of particular interest in view (c) is the dark spot in the upper-left part of the imaged object: this is one of the two 5 cm (two inch) holes in the object that appear in view (a). Here the STIL system is actually ranging down through that hole to the bottom of the object. (The other hole was covered by a weight used to keep the object on the ocean bottom.)
As the scattering and attenuation of water are significantly greater (and propagation velocity significantly smaller) than in air, Areté has developed and tested the algorithms and software to account for such problems. These algorithms are directly translatable to long-range air paths, and propagation through fog, haze, smoke etc.
(b) Safety limitations of conventional lidar—Modern STIL innovations were developed for underwater applications that require blue-green light for optimal water penetration. Human beings too are particularly adapted for sensitivity to light in these wavelengths.
By the same token, however, such light when projected at very high powers can pose a hazard to people—and possibly to other creatures as well—who may be positioned to look directly at the source. The possible hazard is compounded by a like sensitivity to viewing specular reflections of the beam from the source.
As will be understood, the STIL system has many useful applications in which this type of potential hazard poses no significant concern. A thrust of the present document, however, is development of a new generation of STIL systems and applications that are industrial and even commercial, and accordingly introduce a much greater need for compatibility with the population at large.
Therefore the possibility of injury to eyes is an important obstacle to a new array of STIL devices. It may be in part due to this problem that widespread commercial and industrial adaptations of the STIL principle have failed to appear in the marketplace.
(c) Conventional lidar streak-unit limitations—As the preceding introductory sections suggest, conventional modern streak tubes are relatively sizable vacuum tubes that use high voltages to streak the electron beam generated from the photocathode. Plainly this type of hardware is subject to several drawbacks.
Such devices are very expensive to make, maintain and use. For field use, ruggedization is a necessary added expense (and a still-imperfect solution) since large vacuum tubes are inherently somewhat fragile. Their external high-tension connections are not optimal for routine use in aircraft.
A well-known alternative is optical streaking—in which a beam of incoming photons is rapidly displaced across a detector, along the range axis, entirely avoiding the need for a vacuum tube. This in fact was the earliest form of the streak camera—using a fast scan mirror, in particular a large spinning polygon (FIG. 11 [a]).
These devices too, unfortunately, are problematic and even more so than the electronic form. The drawback of this approach as conventionally implemented is the requirement for the large high-speed rotating mirror, which is both bulky and relatively delicate. (One relatively modern example of such an installation is described by Ching C. Lai in “A New Tubeless Nanosecond Streak Camera Based on Optical Deflection and Direct CCD Imaging,”, Proc. SPIE vol. 1801, 1992, pp. 454-69.)
What makes these drawbacks of the optical streaking technique particularly unfortunate is that streaking with an optical device would otherwise greatly expand the choices in commonly available detectors. It would allow the use of common detectors for the wavelengths of interest—e.g. silicon CCD and CMOS detectors for the visible and near IR, and HgCdTe, PtSi, or InSb arrays for the longer IR, out to 11-micron wavelengths if desired.
Longer-wavelength operation would be advantageous for various special applications. These include better penetration of fog, clouds and some types of smoke; and also enhanced discrimination of object types by their different reflectivities at corresponding different wavelengths.
Regrettably the common detectors just mentioned are not suited for use as photocathode materials, to generate electrons that can then be streaked inside a streak tube. On the other hand it would accomplish nothing to place them following the conventional photocathode—e.g. at the streak-tube anode—since conversion from the optical to the electronic domain has already been accomplished at the cathode.
Use of a standard IR imaging detector instead of a CCD would be advantageous to provide high-quantum-efficiency images. For some wavelength ranges this technique would be ideal—but the prior art has avoided these potential solutions because of the recognized problems presented by spinning mirrors.
(d) Conventional lidar imaging limitations—As a general observation, conceptually a STIL system is far in advance of competing technologies in terms of resolution capability in three dimensions, and in terms of signal-to-noise ratio as well. In its ability to fully exploit these advantages, however, a conventional STIL is severely impaired by an overriding problem in streak lidar systems heretofore: inflexibility of pixel allocation.
This limitation may be appreciated from three different perspectives, although in a sense they are only different aspects of the common phenomenon:                the STIL cannot record in three dimensions without mechanical movement of the measuring instrument relative to the region to be inspected;        the only practical way to make optimally efficient use of the very expensive detector area in a conventional STIL system is to build a fiber-optic remapper; and        even when such a device has been built, a conventional STIL system fails to make fully economic use of that investment.These problems will be taken up in turn below, but first beginning with a demonstration of the above preliminary observation that 3-D resolution is superior in a STIL apparatus.        
THREE-DIMENSIONAL RESOLUTION: Different existing lidar systems sample a water volume differently (FIG. 16). The water surface is represented by the irregular line shown on the two visible faces of the volume cube, in each view.
Range-gated systems have excellent transverse spatial resolution, but only have one range pixel per camera—which results in poor range resolution as suggested by the relatively tall volume elements (FIG. 16[a]) in the shaded zone that is of interest. Merely by way of example, one system well-known in this field as “Magic Lantern” is forced to use six separate cameras to cover multiple depths, resulting in a large and expensive system.
In addition, since a range-gated system thus collects large vertical sections of the water column, contrast of any object images is significantly reduced. That is, the contrast, which is directly proportional to the signal-to-noise ratio (SNR) in the region, is a function of the amount of water backscatter that is collected.
A system with range samples of 30 cm (one foot) has ten times the SNR of a system that has 3 m (ten foot) range samples. In addition, as the diagram also suggests, the range-gated device must avoid the surface of the water.
Time-resolved systems, such as the one used in the advanced receiver in ATD-111 (a photomultiplier-tube-based, nonstreaking time-resolved lidar system), suffer from a similar contrast-reduction problem. In this case the cause is poor transverse spatial resolution, as suggested by the relatively broad volume elements (FIG. 16[b]) in the shaded zone of interest.
This system cannot isolate an object signal, and moreover also collects a large area of water backscatter around an object. To have the same transverse spatial resolution as the range-gated system, this time-resolved apparatus would require a separate laser pulse for every pixel, resulting in a pulse repetition frequency (PRF) exceeding 100 kHz.
Unfortunately a frequency-doubled Q-switched Nd:YAG laser (the primary laser used in ocean lidar systems) operates efficiently only up to about 5 kHz. Inability to reach the needed PRF, in turn, results in larger spatial pixels to achieve the same area coverage.
To avoid these sampling problems, the two systems discussed above use both a time-resolved receiver and a range-gated module. Although this approach represents significant additional system complexity, it still does not resolve the significant degradation of detection SNR.
Because a STIL system collects 500 to 1000 spatial pixels per laser pulse, the PRF can be in the hundreds of hertz, which is well within the performance envelope of the Nd:YAG lasers. In this way a STIL device can achieve pixel sizes smaller than an object of interest; therefore, it has higher SNR for the same laser power.
Thus, a streak-tube-based system (FIG. 16[c]) can provide much higher SNR for the same amount of laser power, or can achieve equal performance with a significantly smaller laser system. Streak-tube-based systems can provide good resolution in all dimensions.
Unfortunately this powerful benefit of the STIL principle is not heretofore broadly available without mechanical movement of the detector, and without costly and awkward remapping devices—and even then carries only very limited amounts of image information. These three problems are discussed in the paragraphs below.
THE MECHANICAL-MOVEMENT REQUIREMENT: As described earlier, the conventional streak-tube system is a pushbroom system, which means that it depends on the motion of the vehicle to sample the dimension along the track. This requirement prevents the conventional STIL from serving as what may be called a “staring” system—i.e., a stationary system that can acquire a stationary image of an area.
Just such a capability, however, is quite desirable for a number of useful applications. Inability of a conventional STIL instrument to fill this role is a major limitation in industrial and commercial uses.
FIBER-OPTIC REMAPPERS—CHARACTER AND COST: It is well known to use a variety of kinds of fiber-optic units to reconfigure a time-varying area image as a line image, and thereby enable time resolution of the changing content in the area image. Such technologies are seen in representative patents of Alfano (for example see U.S. Pat. No. 5,142,372), and of Knight (for example U.S. Re. 33,865); and in their technical papers as well.
An original concept for an area-image streak-tube system was demonstrated by Knight, who mapped a 16×16-unit areal image onto a conventional streak-tube slit, with fiber optics. (F. K. Knight, et al., “Three dimensional imaging using a single laser pulse”, Proc. SPIE vol. 1103, 1989, pp. 174-89.) This technique was severely limited in overall number of spatial pixels because of the relatively small number of pixels that can be mapped onto a slit.
Low-resolution fiber image redistribution (a 16×16 focal plane to a single 256-pixel line) has also been performed for streak tubes by MIT-Lincoln Labs. Many fiber-array manufacturers are in operation and ready to prepare units suited for STIL work: one of the largest firms is INCOM; another that makes individual fiber arrays is Polymicro Technologies—which has previously prepared arrays with 3000 fibers.
At best, however, all such approaches are hampered by the costly custom fabrication required, and the need to manufacture a special unit for each desired mapping respectively.
FIBER-OPTIC REMAPPERS—INADEQUATE EXPLOITATION: What makes matters worse, as to fiber-optic remapping, is that a conventional STIL system nowadays continues to face the same basic obstacle seen in the Knight paper noted above. Only so many original image pixels can be meaningfully rearranged onto a slit.
This means that even after the limitation of expensive custom fabrication has been confronted and in a sense overcome by a decision to expend the necessary funds, and even after the requirement for making a separate special unit for each of several particular mappings has also been faced and in a sense overcome by a decision to invest even that multiple—yet nevertheless the technology continues to be not only uneconomic but also technically unsatisfactory because the resulting images carry inadequate, frustratingly small amounts of image information. This obstacle has heretofore remained a persistent problem, and will be further discussed shortly in subsections (f) and (g).
(e) WFS limitations of conventional lidar—The field of wavefront sensors (WFS) is an important one for laser diagnostics. High-power short-pulse lasers are essential components of several different applications (e.g., laser trackers and imaging laser radar); however, such lasers are notoriously unreliable.
It is difficult for vendors to manufacture them to desired specifications, and the devices seldom survive to their projected lifetime (at least at rated output power). One of the most difficult aspects of the manufacture of such devices is the lack of diagnostic equipment for the total characterization of the laser output.
Typical laboratory equipment for the characterization of high-power pulsed lasers consists of three instruments: (1) a power meter for measuring average power, (2) a single fast detector with an oscilloscope for measuring the pulse width temporally, and (3) a laser characterization imager that provides a spatial display of the beam intensity. Each of these instruments has significant limitations in the data that it produces.
The single fast detector averages over the spatial components of the beam, and the laser characterization imager averages over the temporal component of the beam. That is to say, the integration time of the camera in the laser characterization setup is typically orders of magnitude longer than the laser pulse width.
The power detector, furthermore, averages over both the spatial component and the temporal component. Thus no one instrument provides information in time and space and phase with high resolution in all dimensions.
Yet this is precisely the information that the laser designers use in their modeling and simulations. Using commercial laser modeling software such as GLAD, laser designers set up models to simulate propagation of the beam in the laser cavity at very high spatial and temporal resolution.
The phase and intensity of the light, expressed as electric fields, are used in this simulated propagation process. After going through all of that analysis, however, laser designers have no way to compare the actually resulting, operating product with that preliminary analysis.
None of the above-discussed three instruments measures the wavefront of the light—i.e., maps the phase of the outgoing light as a function of position in the beam. This is the role of another type of instrument, the WFS, which does exist to perform this task—but like the laser characterization imager it averages over time.
The most common such unit in use today is the Hartmann-Shack WFS. In this apparatus, incoming light (FIG. 25) is split up into multiple subapertures, each with its own lenslet. The lenslet focuses the light onto a detector.
When a flat wavefront (i.e., a plane wave) is incident on the device, each of the lenslets forms a spot image on-axis on the detector. When a distorted wavefront arrives, however, as illustrated the average slope of the wavefront at the lenslet for each subaperture displaces the spot away from the on-axis position.
Although the illustration is essentially one-dimensional, the lenslets are in a two-dimensional array; and the spot position measurements too are accordingly made in two dimensions. Design and fabrication of this kind of device is a highly specialized endeavor, available from various vendors such as Wavefront Sciences, Inc. of Albuquerque, N. Mex.
Wavefront Sciences develops lenslet arrays for a number of applications. Typical cost for initial design and fabrication of one lenslet array is $20,000.
The displacement of the spot is measured, in both the x and y directions. Average local slope of the wavefront at the measurement point is next calculated as linearly proportional to this displacement. The total wavefront is then reconstructed using algorithms that assemble such local tilts into a whole wavefront.
This process is referred to as “wavefront reconstruction”. It is a common and well-documented algorithm, currently used in astronomical and many other instruments.
In addition to the wavefront, which corresponds to the phase of the electric field, the intensity of the light is measured for each subaperture. This allows generation of an intensity map, as well as a phase map, of the incoming beam.
In most Hartmann-Shack WFS units, the detector behind the optics is a CCD camera or an array of quad cells. A quad cell (FIG. 26) measures the two tilts and the intensity. These detector systems are relatively slow (30 Hz to 10 kHz); while sufficient for assessing atmospheric corrections such detection naturally is inadequate for applications that require subnanosecond sample rates.
From the foregoing it will be clear that laser laboratory devices, and in particular WFS systems when used for laser evaluation, fail to satisfy the needs of laser developers. This failure is a major problem, impeding progress in the design and refinement of more stable, reliable and long-lived lasers.
(f) Data-speed and package limitations—Signal processing in conventional STIL systems is performed using multiple digital signal processors (DSPs). These in turn impose requirements of weight, volume, power, and heat-loading which in effect demand vehicle-mounting of these sensors.
Even carried on a vehicle of modest size, practical forms of the system have relatively low data throughput and may therefore require several measurement passes to acquire adequate data for a region of interest. These limitations represent additional problems because many applications would be better served by a system that a single person could carry, or that could survey and map a region in a single pass—or ideally both.
(g) Limited uses of conventional streak lidar—No STIL packages fully suited for commercial or industrial surveillance and mapping are known to be on the market. It appears that this may be due to a combination of factors including the visual hazards mentioned earlier (with the legal liability that would be associated with operations in populated areas), and also the limited data speed and resulting packaging obstacles outlined just above.
Potentially, a primary commercial application is airborne three-dimensional terrain mapping. Terrestrial mapping is one function that can be performed using lidar, but this opportunity has not been exploited commercially. It is believed that this market may represent potential income exceeding tens of millions of dollars annually.
In California, for example, there is a need to perform complete surveys of the Los Angeles basin (2,400 square miles) every year. This task is currently performed using photogrammetry techniques. Other metropolitan areas have similar requirements, which in the aggregate thus can provide a sustained business in airborne surveying.
An entree to this terrestrial mapping application can be obtained by contacting any large commercial and industrial surveying company. A very roughly equal amount of business can be generated through “on demand” surveying for particular construction jobs—particularly three-dimensional imaging.
Conventional STIL equipment, however, has not been set up (or at least not set up in a convenient format) for three-dimensional imaging. Likewise it is not available with any kind of viewing redundancy, to surmount problems of temporary or local barriers to viewing.
On land such barriers include for example landscaping or natural forestation, as well as coverings deliberately placed over some objects. At sea they include image-distorting effects of ocean waves (FIG. 19)—which may completely obscure some features and actually exchange the apparent positions of others.
In purest principle it is known that foliage and other kinds of cover can be neutralized through use of spectral signatures, polarization signatures or fluorescence signatures. Analysis that incorporates spectral, polarization, spectropolarization, and fluorescence discriminations is also known to be useful for other forms of optical monitoring for which streak lidar would be extremely well suited.
Significant analysis of three-dimensional polarization analysis with lidar systems, using a “Mueller matrix” approach, is in the technical literature. See, for example, A. D. Gleckler, A. Gelbart, J. M. Bowden, “Multispectral and hyperspectral 3D imaging lidar based upon the multiple slit streak tube imaging lidar”, Proc. SPIE vol. 4377, April 2001; A. D. Gleckler, A. Gelbart, “Three-dimensional imaging polarimetry”, Proc. SPIE vol. 4377, April 2001; A. D. Gleckler, “Multiple-Slit Streak Tube Imaging Lidar (MS-STIL) Applications,” Proc. SPIE vol. 4035, p. 266-278, 2000; R. M. A. Azzam and N. M. Bashara, Ellipsometry and Polarized Light, North Holland, Amsterdam (1977); R. A. Chipman, E. A. Sornsin, and J. L. Pezzaniti, “Mueller matrix imaging polarimetry: An overview” in Polarization Analysis and Applications to Device Technology, SPIE Volume 2873, June 1996; R. M. A. Azzam, “Mueller matrix ellipsometry: a review”, SPIE Volume 3121, August 1997; P. Elies, et al., Surface rugosity and polarimetric analysis, SPIE Volume 2782, September 1996; Shih-Yau Lu & R. A. Chipman, Interpretation of Mueller matrices based on polar decomposition, JOSA, Volume 13, No. 5, May 1996; and S. Bruegnot and P. Clemenceau, “Modeling and performance of a polarization active imager at λ=806 nm”, SPIE Vol. 3707, April 1999.
Thus spectral, fluorescence and polarization analyses in theory are susceptible to commercial and industrial streak-lidar beneficial exploitation. Examples are detection and measurement of atmospheric particulates, atmospheric constituents, waterborne particulates, and certain hard-body object returns (with propagation paths in either air or water).
Heretofore, however, necessary equipment adaptations for introducing fluorescence, polarization and spectral analyses into streak lidar work—at least on a broad, general-use basis—have been unavailable. The prior art in this field thus fails to teach how to go about making such refinements in any straightforward, practical way. This gap represents a major problem as it has left these kinds of mapping infeasible or even impossible—and accordingly several practical mapping needs unsolved.
(h) Now-unrelated technology: “eye safe”—This discussion will next turn to modern developments that have not heretofore been pragmatically associated with lidar, or particularly with streak imaging lidar. The first of these relates to population exposure.
Studies have shown that light beams of different wavelengths have respective ocular destructive powers—for any given beam power—that differ by many orders of magnitude. For all wavelengths, such studies have established respective maximum power/pulse-energy levels that are considered safe, at least for humans.
In particular the 1.5-micron region is considered to have the least ocular destructive power (by several orders of magnitude) of any wavelength from x-rays to the far infrared. More specifically, light in the visible, near-UV, and near-IR regions damages the retina, while light in the far-UV and far-IR damages the cornea; but light at about 1.5 microns tends to dissipate harmlessly in the intraocular fluid between the cornea and the retina. This region is therefore commonly designated “eye safe”.
Accordingly for mapping or detecting systems that are to irradiate large areas of land in which people or other higher organisms may be present, it is important to operate in that eye-safe wavelength region as much as possible. Of course there are many reasons to avoid incidental exposure of people above the damage threshold.
Traditional photocathode materials are well suited at visible wavelengths; however, efficient photocathode detector materials do not exist for wavelengths much over one micron. Nevertheless operation at eye-safe wavelengths is feasible with commercially available fast phosphorescent materials, which respond to infrared photons by producing proportional quantities of higher-frequency (visible) photons, and are thus loosely described as performing wavelength “conversion”.
These materials thus in effect “convert” light at 1.5 microns to roughly 0.65 micron, with quantum efficiencies potentially as high as sixty-six percent. This process, sometimes called “upconversion” to visible light, occurs at the front of an imaging tube—in advance of the photocathode—and enables use of conventional photocathode materials that respond well to visible light.
TRANSMITTER: Transmitters operating at 1.5 microns are now commonly available from several vendors (including Big Sky Laser, LiteCycles, and GEC-Marconi) using diode-pumped solid state (DPSS) Q-switched Nd:YAG lasers. The lasers are coupled with either an optical parametric oscillator (OPO) or a stimulated Raman scattering (SRS) cell.
To achieve optimal range resolution, it is desired to keep the laser pulse length between 4 and 10 nsec. This is the range of typical DPSS Q-switched pulse lengths; accordingly transmitter conversion is straightforward within the state of the art.
RECEIVER: Streak-camera receiver operation at 1.5 microns is currently available using the standard S1 photocathode material. Unfortunately, S1 has very poor quantum efficiency, which reduces the applicability for real-world imaging.
Due to the low efficiency of the S1 material, a 1.5-micron streak-lidar product based upon it would similarly operate inefficiently. Pragmatically speaking, such a product would not be economic or viable.
Exploration of other materials has been reported. Those materials which are relevant to vacuum streak-tube operation include:                TE photocathode,        InGaAs photocathode,        ETIR upconversion, and        nonlinear upconversion.These will be discussed in turn below, in this subsection of the present document. Another approach to eye-safe technology, but that excludes vacuum streak-tube operation entirely, will be discussed in a later subsection.        
TE Photocathode: Intevac Corp. has fabricated a transfer-electron (TE) photocathode with quantum efficiencies exceeding ten percent at 1.5 micron wavelengths. (See K. Costello, V. Abbe, et al., “Transferred electron photocathode with greater than 20% quantum efficiency beyond 1 micron”, Proc. SPIE vol. 2550, pp. 177-88, 1995.) This photocathode has been demonstrated in image-intensified CCDs—but not in streak tubes—at 1.5 microns.
Very interestingly, applicability of the TE photocathode for streak tubes has been shown too, but not at that eye-safe wavelength. (Please refer to V. W. Abbe, K. Costello, G. Davis, R. Weiss, “Photocathode development for a 1300 nm streak tube”, Proc. SPIE vol. 2022, 1993.)
Intevac used an early version of this photocathode in a streak tube that operated out to 1.3 microns—but, again, not to 1.5. Whether actually due to perceived lack of customer base or due to some failure in reduction to practice, no streak tube able to operate efficiently at 1.5 microns is currently available.
InGaAs Photocathode: Hamamatsu Corporation of Japan has an InGaAs photocathode used for near-IR photomultiplier tubes (PMTs). The Hamamatsu photocathode has poorer quantum efficiency (QE)—on the order of one percent—at 1.5 microns than the TE discussed above.
Although Hamamatsu suggests that this photocathode as compatible with its streak-tube line, no such development has appeared, at least commercially or in the literature. Again, pragmatically no successful report of testing is known.
Improvements in QE may be possible if “slower” photocathodes that have a longer response time—1 nsec, vs. tens of picoseconds—are acceptable. Nanosecond response time at the photocathode would have little adverse impact on system performance.
ETIR Phosphor Upconversion: Phosphor upconversion has been performed by simply placing a layer of phosphor in front of a photocathode, in image intensifiers and other photoresponsive devices. No testing with a streak-tube photocathode, however, has been reported.
In known applications of the phosphor-upconversion technique, the incoming IR interacts with the phosphor, which has been “charged” with blue light from an LED, and in response produces light between 600 and 700 nm. This is well within the high-performance range of conventional photocathode materials. The blue charging LED can be shut off during the brief data-collection period to avoid saturating the photocathode.
This technique is particularly effective using a class of phosphors, called “electron-trapping infrared” (ETIR) upconversion phosphors, which receive incident infrared photons and in response emit corresponding quantities of visible photons.
The response band of typical ETIR phosphors is about 0.8 to 1.6 μm. The most accepted model for the operation of ETIR phosphors is as follows.    (1) The phosphors are doped such that there are two doping levels between the valence and conduction bands, with the lower doping level called the “trapping level” and the upper doping level called the “communication level”.    (2) Visible photons (typically blue to green) excite electrons from the ground state to levels higher than the trapping level.    (3) In the combination process, most electrons then fall to the trapping level where they can remain for very protracted time periods (years), in the absence of infrared photons with energies corresponding to the gap between the trapping and communication level.    (4) Incident infrared photons excite the electrons in the trapping level to the communication level, where they radiatively decay to the ground state by combining with holes in the ground state—releasing visible photons (typically orange to red).
Another class of infrared upconversion phosphors is anti-Stokes (AS) phosphors. Since these phosphors operate via a multiphoton process, they have higher thresholds and lower conversion efficiencies than do the ETIR phosphors. AS phosphors, however, do not require visible pump photons for operation as do the ETIR phosphors. The need for a visible pump is not a major drawback for the ETIR phosphors, since the pump need not be coherent—and hence LEDs can be used as the pump source.
Nonlinear optical processes competing with ETIR phosphors include second harmonic generation (SHG), stimulated Raman scattering (SRS) antiStokes (AS) lines, and sum-frequency generation. The ETIR phosphors have an advantage over all these competitors, namely that there are no phase matching or coherency requirements, so that the ETIR process can operate over the wide incidence angles required for imaging.
Commercially available ETIR films from Lumitek International, Inc. (formerly Quantex) have been reported with 2 nsec pulse response width and twenty-two percent quantum efficiency (in reflective mode), from 1.06 μm to visible for the company's Q-11-R film. (Ping, Gong and Hou Xun, “A New Material Applicable in the Infrared Streak Camera,” Chinese Journal of Infrared and Millimeter Waves, vol. 14, No. 2, 1996, pp. 181-82.)
Quantex has developed a near-infrared image intensifier (model I2) using an ETIR phosphor screen. In this project the company measured the minimum sensitivity—at several wavelengths—of a thick-film phosphor screen mated with the image intensifier in transmissive mode. (Lindmayer, Joseph and David McGuire, “An Extended Range Near-Infrared Image Intensifier,” Electron Tubes and Image Intensifiers, [ed.] Illes P. Csorba, Proc. SPIE vol. 1243, 1990, pp. 107-13.)
Measured minimum sensitivity of this phosphor-I2 sensor at 1.55 μm was 670 nW/cm2 (ibid. at 108). Quantex had also vapor-deposited thin films of the ETIR material onto an image-intensifier fiber-optic faceplate to improve the imaging resolution. The resolution obtained with this device was 36 line-pairs/mm, corresponding to the resolution of the 15 μm fiber pitch of the faceplate.
Researchers at Xi'an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, have reported the development of faster, more efficient ETIR phosphors. (Ping, Gong and Hou Xun, loc. cit.)
They reported a red-emitting and a blue-emitting ETIR phosphor—the red phosphor having a 1.3 nsec response width and a 66% quantum efficiency in transmissive mode, and the blue phosphor having a 1.4 nsec response width and a 47% quantum efficiency in transmissive mode.
Nonlinear upconversion: Because conversion efficiency is a function of optical power, nonlinear upconversion techniques such as SHG and SRS are not practical for low-level signals. Consequently, these techniques are typically used with the transmitter rather than the receiver.
Also, because of phase matching requirements these techniques are typically only efficient over limited fields of view. There is a technique in which the signal can be amplified optically in a Raman crystal, to allow for efficient upconversion—or to directly overcome the poor quantum efficiency of an S1 photocathode (Calmes, Lonnie C., et al., “Marine Raman Image Amplification”, Proc. SPIE Vol. 3761, 1999). At present a drawback of this technique, with respect to long-range streak-tube operations, is that the Raman amplifier can be pulsed on for only 10 to 15 nsec at a time.
As to operation at longer wavelengths (1.6 to 10 microns), advantageous for various specialized applications as mentioned earlier, Quantex has also reported ETIR phosphors for upconverting medium-wavelength infrared (3.1 to 4.5 μm) to 633 nm light. (Soltani, Peter K., Gregory Pierce, George M. Storti, and Charles Y. Wrigley, “New Medium Wave Infrared Stimulable Phosphor for Image Intensifier Applications,” [ed.] Illes P. Csorba, Proc. SPIE vol. 1243, 1990.) Unlike the eye-safe wavelength phosphors, these require cryogenic hardware for the phosphor upconversion plane.
RECEIVER CHOICE: As can now be seen, a great variety of technology is available for receiving eye-safe radiation and causing phosphor-responsive tube devices to respond. No demonstration, however, of pragmatically efficient operation in a lidar streak tube has been reported. Equipment adaptations accordingly have not been developed.
The foregoing discussions have noted the availability and use of very inefficient S1 detector material at 1.5 microns, and Hamamatsu's views as to its own low-efficiency detector material at that wavelength—both these materials being inadequate for industrial-quality instrumentation in the present state of the art—and also the Abbe experiments with TE material at 1.3 microns, and the suggestion by Ping of using ETIR material in streak cameras. In the absence of dispositive testing, none of these appears to represent an enabling disclosure of a commercially feasible eye-safe STIL system.
(i) Now-unrelated technologies: modern optical deflectors—Another area of technological advances that are known but have not heretofore been connected with streak lidar is microelectromechanical systems (MEMS). These devices are very small, and enable use of a simplified optical path (FIG. 11[b]).
A prominent example is a Texas Instruments product denominated a “Digital Micromirror Device” (DMD™). TI makes its DMD units for the commercial projection display market; accordingly they are readily available.
Key factors for efficient use of a DMD component include the mirror fill factor, scanning speed, uniformity of mirror motions, and quantification of diffraction effects. The DMD product has a fill factor higher than ninety percent, and can scan forty degrees in two microseconds (see Larry J. Hornbeck, “Digital Light Processing for high-brightness high-resolution applications”, A presentation for Electronic Imaging, EI '97, Projection Displays, February 1997).
They are compatible with operation at virtually any wavelength of interest for imaging and detection—including, in particular, the eye-safe technology discussed in the preceding subsection. Again, although well established these devices have not been associated with lidar instrumentation heretofore.
As can now be seen, the related art remains subject to significant problems, and the efforts outlined above—although praiseworthy—have left room for considerable refinement.